The proposed research will use the Andersen model of health care utilization to identify determinants of health care utilization in a sample of 290 elderly Baltimore Jews. The dependent variables will be self reports of physician visits and hospitalization. The Andersen model uses three sets of variables to predict utilization; they are predisposing, enabling, and illness level characteristics. Predisposing variables exist prior to a specific illness episode and include demographic, social structural, and social psychological factors. Enabling variables measure the availability and type of care available to an individual. Illness level characteristics include both perceived and evaluated need. Multivariate techniques will be used throughout this analysis. Linear models will be compared to logarithmic transformations. Proceedures developed by Box and Cox (1963) will be used to assess th appropriateness of the linear model in comparison to the logarithmic formulation. This project analyzes secondary data. The original data were collected for a study on the health, housing, and social service needs of elderly Baltimore Jews. The data were collected during the Summer of 1981.